Job search with Codex

Drimva lists about 130,000 open US full-time jobs and internships from the career sites of more than 2,000 companies, and checks every job each day. Give Codex one instruction and it reads Drimva's workflow, finds the jobs that fit your resume and prepares each application with you.

Start with one instruction

Paste this into Codex CLI or the Codex app:

Read https://drimva.com/llms.txt and follow its job-search workflow to help me find a job.

Optional: install the skill

Codex loads your personal skills from ~/.agents/skills. With the Drimva skill there, you can simply ask Codex for help with your job search. The skill points Codex to the same workflow, so it stays current.

mkdir -p ~/.agents/skills/drimva && curl -fsSL https://raw.githubusercontent.com/JonnesLin/drimva-skill/main/skills/drimva/SKILL.md -o ~/.agents/skills/drimva/SKILL.md

The skill is open source on GitHub.

What Codex does

Reads your resume and picks search terms from Drimva's keyword lists.
Scores each job from 0 to 100 against your background. Jobs whose posting states a condition you don't meet, such as visa sponsorship or a security clearance, are dropped.
Asks your preferences, such as places and minimum pay, and shows how many jobs each answer keeps.
Gives you a shortlist. You pick the jobs.
Prepares each application. It asks only for answers it doesn't have and, if you want, tailors your resume to the job. You see each change before it is used.
Fills in the forms when Codex has browser tools. Without them, it gives you the answers and files to use yourself.

Things to ask

“Find new-grad software engineering jobs in New York.”
“Find data science internships and rank them against my resume.”
“Prepare applications for my five best matches.”

Your account and your data

Searching needs no account. With a free Drimva account, Codex saves your shortlist and application progress, and you can review both on Drimva. Your resume and application files stay in your Codex workspace, and Codex's own data practices apply.

A short video and the full workflow are on For AI agents. Other guides: Claude Code · MCP server.